I will not address the EJS part in this answer as I'm not qualified and the code you provided seems all good. However, I'll review the back-end part.
Also, since I don't know what kind of coding background you have (if any), this answer will contain a lot of explanation on perhaps simple concepts.
Summary
From your second code snippet, there are a couple things that are to be discussed:
- Asynchronous code
- The database connection and generalities
- Actual implementation
- Code conception
- [EDIT]: Save/Edit implementations
There is also a lot more to cover depending on the knowledge of OP, such as try/catch clauses, MongoDB models validation, the usage of express's Router and more but I will only edit my answer if needed.
Asynchronous code
For the rest of the answer, most of the code will be surrounded by async/await keywords. These are necessary for the code to work properly.
Basically, JS being a language that is made for the web, you sometimes need to wait for network or database requests to be done before you do any other action. That's where the callbacks, the promises or the async/await syntax (which is syntactic sugar for promises) come in handy.
Let's say you need, like your example, to retrieve a list of tasks:
app.route("/:list_name").get((req, res) => {
MongoClient.connect(process.env.DB_CONNECT, (err, db) => {
if(err) throw err;
var database = db.db("myFirstDatabase");
const cursor = database.collection(req.params.list_name).find({}); /* stuck here */
console.log(cursor);
// ..........
});
});
JS being asynchronous by default, if you run this code, chances are high that cursor will be undefined. The reason for that is that the code doesn't wait for the database.collection(............. to finish in order to continue the execution. But with the help of the aforementioned callback/promises/async-await, our code can now wait until this instruction is done.
You can read on async/await here and here, and see here that MongoDB examples are using async/await as well, but you will see in the following sections more "practical" usages of it.
Keep in mind that what you are using (whether it is callbacks, promises or async/await syntax) is completely up to you and your preferences.
Database connection
As the code is currently written, everytime a user clicks on any item on your list, a connection to MongoDB will be established, and that connection doesn't belong to the route handler. Your back-end app should connect to the database once (at least for this case, it could prove useful to initiate multiple connections for some advanced cases), and close the connection when your back-end app stops (generally not the case with an API).
Atlas cloud databases for example, have a limit of 500 connections. Meaning that if, let's say, 501 users click on an item simultaneously on your front-end list, the best case scenario is someone doesn't get what he asked, but it could be worse.
For this matter, you have several options. One would be to go with a framework that helps you leverage some of the code and boilerplate, such as Mongoose or work with the native MongoDB driver which we will do, since you seem to already work with that and I strongly believe working with the lowest layer first will make you learn higher-level frameworks way faster.
Now, let's tackle the problem. We want to put the database connection somewhere else where it'll be called once. Again, there's several options you can go with, but I like to create a class for it, and exporting a new instance to do what I want anywhere in my code. Here is a (really) simple example of what my minimal go-to looks like:
mongo-client.js:
const { MongoClient } = require('mongodb');
class MongoCli {
constructor() {
let url = `mongodb://testuser:[email protected]:27017/?authSource=my_database_name`;
this.client = new MongoClient(url, { useUnifiedTopology: true });
}
async init() {
if (this.client) {
await this.client.connect();
this.db = this.client.db('test');
} else
console.warn("Client is not initialized properly");
}
}
module.exports = new MongoCli();
Actual implementation
Of course, this code on his own won't work and we need to call and wait for it, before defining routes. So, right before app.route("/:list_name")............, call this: await MongoCli.init();.
Here is what my (again, really) simple server.js look like (I have separated the mongo-client code from the server):
const express = require('express');
const MongoCli = require('./mongo-cli.js');
const server = async () => {
const app = express();
await MongoCli.init();
app.route("/:list_name").get(async (req, res) => {
});
return app;
};
module.exports = server;
Now, let's start implementing what you really want from the beginning, a.k.a once a user click on a topic of tasks, it will display all the tasks on the topic he clicked:
const express = require('express');
const MongoCli = require('./mongo-cli.js');
const server = async () => {
const app = express();
await MongoCli.init();
app.route("/:list_name").get(async (req, res) => {
// we will query the collection specified by req.params.list_name
// then, .find({}) indicates we want all the results (empty filter)
// finally, we call .toArray() to transform a Cursor to a human-readable array
const tasks = await MongoCli.db.collection(req.params.list_name).find({}).toArray();
// making sure we got what we needed, you can remove the line below
console.log(tasks);
// return a HTTP 200 status code, along with the results we just queried
res.status(200).json(tasks);
});
return app;
};
module.exports = server;
Quite simple, right?
Keep in mind my server.js might not look quite as yours since there are many ways to handle this and it is to the developer to find his own preferred method, but you get the idea.
Code conception
We got our GET route going, we get the results when we call the route, everything's great! ... not quite.
What happens now if we have, say, 1500 topics of tasks? Should we really create 1500 different collections, knowing that a task consist of a description, a status, a deadline, eventually a name? Sure, we can do it, but it doesn't mean we have to.
Instead, what about creating one and only collection tasks, and adding a key topic to it?
Considering the above sentences, here's what the route would now look like:
const express = require('express');
const MongoCli = require('./mongo-cli.js');
const server = async () => {
const app = express();
await MongoCli.init();
app.route("/:topic_wanted").get(async (req, res) => {
// we now know the collection is named 'tasks'
// then, .find({topic: req.params.topic_wanted}) indicates we want all the results where the key 'topic' corresponds to req.params.topic_wanted
// finally, we call .toArray() to transform a Cursor to a human-readable array
const tasks = await MongoCli.db.collection('tasks').find({topic: req.params.topic_wanted}).toArray();
// making sure we got what we needed
console.log(tasks);
// return a HTTP 200 OK, along with the results we just queried
res.status(200).json(tasks);
});
return app;
};
module.exports = server;
Last words
I hope I'm not too off-topic and my answer could help you.
Also, I saw while writing the answer that you need to figure out how to post tasks now. Please let me know in the comments if you need further information/explanation or even help for posting tasks.
EDIT (added):
Save/Edit implementations
Seeing your implementation of creating a new task, I assume you already use mongoose. Unfortunately, when declaring a model in Mongoose, it will automatically search for (or create if it doesn't exist) the collection having the same name of your declared model, except in lowercase and pluralized (see here for more info). Meaning you can't declare a new task and assign it to a collection named "users" for example.
That's where the part 4 of this answer, "Code conception", comes into play. Otherwise, the code you edited-in has no "major" flaw.